ML20199F741

Safety Evaluation Rept Accepting Design of torus-to-reactor Bldg Vacuum Breakers at Plant

ML20199F741
Person / Time
Site:	Cooper
Issue date:	08/14/1997
From:	NRC (Affiliation Not Assigned)
To:
Shared Package
ML20199F660	List: ML20199F655 ML20199F691 ML20199F741
References
NUDOCS 9802040095
Download: ML20199F741 (9)

Text

m . -

p%

! r p \ UNITSO STATSS 1 8 NUCLEAR REOULATCRY O2MMISSIEN  !

l ,, wAeNMeToN, D.o. asseHcM

{

l i

j STAFF EVALUATION BY THE OFFICE OF NUCLEAR REACTOR REGULATION  !

l.

IN RESPONSE 10 REGION IV TASK INTERFACE AGREEMENT 95-010 l NEBRASKA PUBLIC POWER DISTRICT l 1

C00PER NUCLEAR STAU gi l DOCKET No. 50-298 2 ,

l

1.0 INTRODUCTION

. Region IV Task Interface Agreement (TIA)95-010, dated August 11, 1995, i

< requested that the Office of Nuclear Reactor de ulation (NRR) evaluate the  !

acceptability of the Cooper Nuclear Station

! configuration between the torus (part of the(CN ) isolat90n pr mary valve containment) and the a reactor building (secondary containment). This issue was raised for CNS in j Inspection Report 50-298/95-01, dated March 17, 1995,

2.0 BACKGROUND

\

The configuration at CNS is similar to that for many older boiling water [

i reactors (BWRs) with Mark I containments. The design consists of two

! redundant vacuum relief lines from the reactor building to the suppression

! chamber (torus), each containing two valves in series' an air-operated 3 i

• Thelinesarenominally20inchesin) diameter. butterfly valve, 14CV). PC-A0V-243 The purpose of these

the associated vacuum breakers is to limit a vacuum in the containment.- i Because the lines-penetrate primary containment, the vacuum breakers serve a i

dual _ function

vacuum relief end containmant isolation. The air-operated i butterfly valves are normally closed and are designed to open upon a ,

j. differer,tial pressure of 0.5 psid between the reactor building and the -

4

. suppression chamber. The air-operated butterfly valves have %en designed to l_ fail opea upon loss of air or AC power. This-is the safe position for vacuum relief. Therefore, given an event during which the air supply or the electric i power cannot be assumed to be operable, the air-operated valve cannot be '

4 relied upon for containment isolation, and the single check valve-in each line

.must perform the containment isolation safety function. In general, the air supply and flie electric power supply to the butterfly valves are not safety- '

related, although they have been upgraded in some older BdRs and designed as

- safety-related in some of the more recently-licensed BWRs. The licensee for Cooper has not upgraded the motive power for these valves. Both valves are

, listed in the CNS Updated Safety Analysis Report (USAR) as containment -

4 isolation valves. The valves are stroke tested in accordance with the

! American Society of Mechanical Engineers (ASME) Code once per cycle to ensure. .

i -that they will opon with a 0.5 psid differential pressure and reclose. The c valves are also leak tested as containment isolation valves in accordance with 10 CFR Part 50 Appendix J.

i 9902040095 990130 PDR ADOCK 05000298 P PDR ,

i ;~.-. _ . 1 _ - _ .. . _ _ _ _ _ _ _ . . _ _ _ . . - . _ . _ _ ~ . - . _ _ _ . .

2-The issue raised in TIA 95-010 is-the acceptabi1 My of the containment isolation function in which the check valve is the only containment isolation valve if the nonsafety grade motive power to the butterfly valve is not available.

This issue was first raised as a result of reviews requested in NRC Generic Letter 88-14, " Instrument Air Supply System Problems Affecting Safety-Related Equipment," which was issued or. August 8, 1988. In part, the generic letter requested each licensee to verify that air-operated safety-related components ,

would perform as expected during all design basis events. Several licensees identified the torus-to-reactor building vacuum breakers as a concern and took various actions to address the issue. The February 10, 1989, response from the Nebraska Public Power District for CNS stated that the procedures for the suppression chamber-to-reactor building vacuum breaker fenctional test would be upgraded. No more information was provided.

A series of letters was exchanged between the NRC and the Vermont Yankee Nuclear Power Company:(VYNPC), the-licensee for the Vermont Yankee Nuclear Power Plant. That licensee took the position that This situation is acceptable as this configuration allows both safety functions of this line to be met, using a highly reliable, .... passive check valve to accomplish the isolation function. Vemont Yankee believes that this was the intent of the original designers and the reasen for the special considerathn given this penetration. Therefore, these valves are in compliance with the current licensing bases, (letter from VYNPC to U.S. Nuclear Regulatory Commission (NRC), dated February 15,1990).

The NRC staff disagreed with this position. In a letter to VYNPC dated August 15, 1991, the NRC staff stated that:

...due to the greater importance of the vacuum relief function... the fail safe position of the air-operated valves should be to fail open upon loss of instrument air. The fail open position would ensure vacuum reilef function through the check valve during negative pressure scenarios and provide a containment isolation function through the check valve during positive pressure scenarios.

However,

... considering the importance of ensured containment isolation, the air supply system to the [ air-operated butterfly) valves should be safety grade to minimize the probability of loss of the containment isolation function and all efforts should be placed on preventing failure of the air supply system to the air-operated valves. The Technical Specifications (TSs), Final Safety Analysis Report, and original General Electric Design Documents all indicate that both the check valve and the air-operated valve are containment isolation valves. As containment 4

, sm-. asmi.r .

isolation valves the licensee should ensure that the valves are fully qualified and provided with safety grade electrical and air supplies, where used for control or motive force.

Before th'is letter was sent to VYNPC, the Boiling Water Reactor Owners Group (BWROG), by letter dated July 1,.1991, in response to previous correspondence between YYNPC and the NRC, stated that since the NRC staff's position could potentially affect several other plants, a committee had been formed to tuvcstigate the issue. For several years, the NRC had also been discussing this issue with the licensees for Millstone Unit I and Oyster Creek. At 0yster Creek this issue had been identified as a followup item to an NRC Diagnostic Evaluation Team inspection.

The BWROG's position was communicated to the NRC by letter dated January 24,

.1992. The letter listed the utilities that participated in the development of the position, including the Nebraska Public Power District, the 11ccnsee for CNS. The BWROG's position was similar to that of VYNPC. The BWROG put considerable emphasis on the reliability of the check valve and its classification as a passive component rather than an active component. The significance of this classification is that, "a passive failure (along with loss- of air-operated valve-(A0V) operability) would be necessary to prevent irolation of either of the vacuum relief lines." SECY-77-439, " Single Failure' Criterion" (an NRC information report prepared at the request of the Commission), considers a check valve to be a passive component and therefore, failure of a simple check valve to move to its correct position when required is an example of a passive failure of a fluid system. SECY-77-439 P.lso states,accordingtotheBWROGletter,thatitisnotnecessarytoysume single passive failures in fluid systems such as vacuum relief lines . The BWROG 1etter is the most complete statement of the industry's position on this issue. The NRC staff has never replied to this letter.

3.0 EVALUATION The current requirements applir.able to containment isolation valves are described in Standard Review Plan (SRP) Section 6.2.4, " Containment Isolation System."

General Design Criterion (GDC) 54, " Piping Systems Penetrating Containment,"

requires that piping systems penetrating containment shall be provided with redundant isolation capabilities "which reflect the importance to safety of isolating these piping systems.*

' The statement in SECY 77-439 is actually not quite this emphatic. SECY 77-439 states that "in most instances" the probability of passive failures is sufficiently small that they do not have to be assumed in addition to the initiating failure. This is based on data collected up to 1969.

9 I

4-GDC 56, " Primary Contairment Isolation," specifies acceptable isolation valve configurations for lines that penetrate primary containment and connect directly with the containment atmosphere. It also states that a simple check valve may not, as in this case, be used as an automatic containment isolation.

valve outside containment for a line which is directly connected to the containment atmosphere.

In addition, GDC 56. states that an automatic isolation valve, upon loss of actuating power, shall take the position that provides greater safety. The air-operated valves do open upon loss of actuating power, which is the position of greater safety in this case; the vacuum relief-function of the flow path is considered to be more important to safety than the containment -

isolation function. In addition, there are no regulatory requirements or guidance (including industry standards on containment isolation) which specify that a valve which has taken the position of greater safety must be capable of taking another position during the course of an accident,

, Of those BWRs listed in the January 24, 1992, BWROG letter-(listed in the accompanying table), including CNS, only the Edwin 1. Hatch Nuclear Plant, Unit 2 (Hatch Unit 2) is required to explicitly comply with the General Design Criteria (GDC) of Appendix A of 10 CFR Part 50. Commission guidance in the Staff Requirrnents Memorandum on SECY-92-223. " Resolution of Deviations Identified During the Systematic Evaluation Program," dated September 18, 1992, approves the option presented in SECY-92-223, whereby the staff will generally not apply the GDC to plants with construction permits issued prior to May 21, 1971. Thus, the requirements of GDC 54, " Piping Systems Penetrating Containment," and GDC 56, " Primary Containment Isolation," are not requirements for these plants and they do not need specific exemptions from these GDCs for the design of the containment isolation features for the torus- '

to-reactor building vacuum relief lines. Since the construction permit for Hatch, Unit 2 was issued after May 21, 1971,- the situation for this plant is different and will be discussed below.

The CNS USAR, in Section 2.3.5.1, specifies that there will be two isolation valves, in series, on each Class B line. A Class B valve is defined in the USAR as a valve on a process line that does not directly communicate with the reactor vessel but which penetrates the primary containment and communicates with the primary containment free space; in other words, a containment isolation valve to which GDC 56 would apply in a plant licensed to the present General Design Criteria. However, Table VII-3-1 of the CNS USAR classifies these valves as Class B-X, where the X denotes a variation to the

-classification. In this case, according to USAR Section 1.7.1; " Single Failure Criteria," item 1:

All air-operated valves in the primary containment isolation systems will fail closed upon loss of instrument air supply, with the exception of the Reactor Building to Torus Vacuum Breaker which will fail open upon loss of instrument air.

~ . _ _ . _ _ _ . . _ _ .

l Thus, the actual design is consistent with the CNS USAR.

The January 24, 1992, BWROG 1etter to the NRC staff discusses the design basis for the torus-to-reactor building vacuum relief lines. The letter states that two General Electric Company des'gn specifications apply to the BWRs covered by the letter. General Electric Company Design Specifications 22A1265 and 22A1126 for containment isolation both state that suppression chamber vacuum relief lines utilize 'self-actuated and power operated valves in series" as- -

containment isolation valves. Thus, the design specification requires that the air 4 operated butterfly valves (PC-A0V-243/244-AV) be containment isolation l valves.- This is consistent with the actual-design at CNS and the other-BWRs listed in the BWROG January 24, 1992, letter. These valves are considered as containment isolation valves that fail open as the safe position on loss of motive power. They are-also considered containment isolation valves for purposes of testing and TSs operability requirements.

The CNS Operability' Evaluation on this issue, OE 95-000-013, dated January 31,.

1995, also discussed the o>erating experience with the check valves.- The licensee stated that the cieck valves have never failed to meet-the acceptance criteria and no maintenance has ever been necessary to enable the valves to pass the leak test. VYNPC, in a February 15, 1990, letter to the NRC, made a similar statement. The CNS Operability Evaluation discusses an event at- CNS in which the torus became subatmospheric and a vacuum breaker opened and reclosed. A subsequent Appendix J Type C test was passed. The discussion indicates that although the check valve did not close tightly, the overall Appendix J limit for Type B and C tests was not exceeded.

It should be noted that although the purpose of the lines in question is to perform the primary containment vacuum relief-safety function, it is not expected that this function would be necessary during a LOCA, that is, during ,

the design basis event that requires containment isolation. But, on the other hand, it cannot be ruled out. Revision 4 of the BWR Emergency Procedure Guidelines contains cautionary notes telling the operator to terminate suppression pool spray when the pressure drops below 2.0 psig. Failure of the operator to act could result in a vacuum.

CNS Operability Evaluation 95-000-013 states that:

In the present LOCA analysis, as described in the USAR...the con +ainment pressure drops below 20 psig in less than one hour but never drops below atmospheric pressure. Thus, in the existing analysis there is never a

. need for the check valve in the relief line to open... .

However, as the CNS Operability Evaluation points out, this analysis was done with assumptions intended to maximize containment pressure. The licensee has not analyzed tHs case with assumptions which would minimize the containment pressure.

____.m_m__ _ _ - - - - - - . - -

i There are circumstances in which there may be a vacuum during a LOCA. For instance, there may be a vacuum if a LOCA occurs while the containment is being vented. Noncondensable gases would be expelled and the steam atmosphere would condense. (Since Mark I containments are inerted during operation, technical specifications allow venting only within 24 hours2.777778e-4 days <br />0.00667 hours <br />3.968254e-5 weeks <br />9.132e-6 months <br /> of startup or.

shutdown.) CNS Operability Evaluation 95-000-013 also discusses the case in which the suppression pool water is at a temperature of less than 95'F.

Depending on the actual suppression pool temperature, a vacuum in'the containment might occur.

Thus, it is possible that the check valve may be called upon to change position. The BWROG argues that even though this may be the case, SECY-77-439 permits the check valve to be considered a passive component for which a single failure does not have to be assumed. This is evaluated below.

The single failure criterion of 10 CFR Part 50, Appendix A is not applicabP to the plants listed in the BWROG January 24, 1992, letter with the except ..n of Hatch Unit 2. Each of these plants, however, has its own version of this criterion. As discussed above, the vacuum breaker design is consistent with the USAR discussion of the single failure criterion for CNS.

Standard Review Plan Section 6.2.4 states that the General Design Criteria

... in general, required that two isolation [ valves) in series be used to assure that the isolation function is maintained assuming any single active failure in the containment isolation provisions."

SECY-77-439, which was quoted in the BWROG January 24, 1992, letter, in attempting to document the NRC staff's position, defines a passive failure in

.a fluid system as,

... a breach in the fluid pressure boundary or a mechanical failure which adversely affects a flow path. Examples include the failure of a simole check valve to move to its correct oosition..."

Thus, SECY-77-439 considers the failure of a check valve to be a passive failure.

SECY-94-084, " Policy and Technical Issues Associated With the Regulatory Treatment of Non-Safety Systems in Passive Plant Designs," dated March 28, 1994, more correctly reflects the staff's position (underlined for emphasis):

The staff normally treats check valves, exceot for those in containment isolation systems, as passive devices.

Thus, the failure of a check valve in a containment isolation system is considered an active failure. This position was previously documented in 'the standard ANS-56.2/ ANSI Standard N271-1976, " Containment Isolation Provisions

}

for Fluid Systems," approved on June 28, 1978, which is endorsed by Regulatory Guide 1.141 of the same title. ANS-56.2/ ANSI N271-1976 states that an example of an active failure is failure of a check valve to move to its correct position. Thus, whereas SECY-77-439 makes no distinction as to the function of the check valve, other regulatory guidance is clear that the failure of a check valve, when it is a containment isolation valve, is an active failure.

Therefore, the BWROG discussion which considers a containment isolation check valve failure to be a passive failure is not a correct interpretatien of the Commission's or the industry's guidance on this issue.

However, for the BWRs included in the BWROG's January 24, 1992, letter to the NRC staff addressing this issue, these positions have been documented after these plants received their operating licent.as. The licensing bases for the torus-to-reactor building vacuum breakers in these plants are censistent with their design.

Even though General Design Criteria 54 and 56, Standard Review Plan Section 6.2.4 and Regulatory Guide 1.141 are not applicable to the plants covered by the January 24, 1992, BWROG letter, it appears that these plants satisfy these positions with respect to closure of the air-o >erated valves. GDC 54 requires redundant isolation capabilities. The vacuum 3reaker lines are provided with two isolation valves. GDC 56 requires certain valve configurations "unless it can be demonstrated that the containment isolation provisions for a specific class of lines are acceptable on some other defined basis." In this case, the other defined basis is both valves outside containment, which is allowed by Standard Review Plan Section 6.2.4.II.6.d. The air-operated butterfly valve failing in the open position is in compliance with the explicit requirements of GDC 56. The BWROG and the staff have both stated that the vacuum relief function is the more important safety function. Thus, the opening of the air-operated valve on loss of motive power complies with GDC 56. None of the guidance documents specifies a safety-related air or electric power supply to the air-operated butterfly valve.

Treating the cht k valve as an active component and assuming a single failure of the check valve would leave the line open during a LOCA. However, in this case, CNS and the other plants covered by the January 24, 1992, BWROG letter are not required to comply with the single failure criterion of 10 CFR Part 50, Appendix A. This design does comply with the single failure discussion in the CNS USAR.

Thus, the design of the torus-to-reactor building vacuum breakers is consistent with the plant's licensing basis and with present guidance and requirements for the BWRs listed in the BWROG January 24, 1992, letter.

However, this conclusion does not address the safety aspects of this design.

Northeast Utilities, in a June 2, 1989, letter to the staff, presented a risk evaluation of this issue that concluded that the maximum benefit to be gained

f

-B-f.

from improving the design to eliminate the vulnerability to loss of motive power to the air-operated valves would be 4 person-rem over the remaining life of Millstone, Unit 1 (21 years at the time the analysis was done). This is a small improvement. The NRR staff has independently checked this estimate and believes it is reasonable. Using the guidance of NUREG-1530, " Reassessment of NRC's Dollar Per Person-Rem Conversion Factor Policy,*.the resolution of this issue by hardware changes would not be justified in tems of the cost to avoid this potential offsite dose.

Thus, in summary, the licensee's treatment of the torus-to-reactor building containment isolation valves is in compliance with the licensing basis for the CNS. In order to require a change to this licensing basis, that is, a change to the design of the torus-to-reactor building vacuum breaker line isolation valve configuration, a backfit would be required in accordance with 10 CFR 50.109, "Backfitting." A compliance backfit in accordance with 10 CFR 50.109(a)(4 is not appropriate, since the licensee is in compliance with the CNS license)and all applicable regulations and regulatory guidance. Improving the design is not necessary to ensure adequate protection. This is also true of-the other BWRs listed in the January 24, 1992, BWROG 1etter. Therefore, a backfit in accordance with 10 CFR 50.109(a)(3) would be required. The staff would be required to demonstrate a substantial-increase in the overall

- protection of the public health and safety and that the costs of implementation are justified. As discussed above, in qualitative and quantitative terms, such a backfit appears to offer little improvement'in safety. Therefore, although there would be some increase in safety by supplying safety-related motive power to the air-operated valves, the gain in safety does not justify pursuing the change.

Therefore, NRR considers this issue closed for all the BWRs included in the BWROG January 24, 1992, letter, except for Hatch Unit 2, as previously discussed.

Principal Contributor: R. Lobel Date: August 14. 1997 l

i PARTICIPATING UTILITIES IN BWROG RESPONSE TO TORUS-TO-REACTOR BUILDING VACUUM BREAKER ISSUE UTILITY APPLICABLE POWER DATE OF CP*

PLANTS Commonwealth Edison Quad Cities 1 2/15/67 Quad Cities 2 2/15/67 Dresden 2 1/10/66 Dresden 3 10/14/66 GPU-Nuclear Oyster Creek 12/15/64 Nebraska Public Cooper 6/4/68 Power District New York Power FitzPatrick 5/20/70 Authority Northeast Utilities Millstone 1 5/19/66 Northern States Monticello 6/19/67 Power Southern Nuclear Hatch 1 9/30/69 Hatch 2 12/27/72 Vermont Yankee Vemont Yankee 12/11/67 Nuclear Power Corporation

• NOTE: Comission Staff Requirements Memorandum on SECY-92-223, dated September 18, 1992, states that the Comisa. ion approves the o) tion presented in SECY-92-223 in which the staff will not apply tie General Design Criteria (GDC) to plants with construction permits issued prior to May 21, 1971, and therefore, these plants do not need exemptions from the GDC. Note that Hatch Unit 2 is the only plant in the table for which the GDC of 10 CFR 50 Appendix A are applicable.

- _.